Detection of electrocardiographic ‘left ventricular strain’ using neural nets

The use of artificial neural networks for classification of ST-T abnormalities of the electrocardiogram (ECG) was investigated. A training set of 356 lateral leads selected from 105 ECGs was visually classified as exhibiting one particular ST-T morphology (left ventricular (LV) strain) or not. Selected measurements, together with the classification, were fed as input to a three-layer software-based network during the learning process. The performance of the network was evaluated by comparing the results obtained from the network with conventional criteria, using two test sets. Set 1 comprised 63 lateral leads from 32 ECGs with ST-T changes showing atypical forms of LV strain. Set 2 consisted of 80 lateral leads from 20 ECGs containing normal and abnormal T-waves. For set 1, the network outperformed conventional criteria, having a higher sensitivity (96 per cent against 85 per cent) and specificity (67 per cent against 50 per cent). With test set 2, both network and conventional criteria were 100 per cent sensitive and 100 per cent specific. For sets 1 and 2 combined, the network had a higher overall sensitivity (97 per cent agaisst 89 per cent) and specificity (88 per cent against 82 per cent). The results suggest that neural networks may be useful in selected areas of electrocardiography, but care is required when selecting patterns for use in the training process.     

微晶蜡非催化氧化反应的神经网络模型

对微晶蜡进行了非催化空气氧化，并利用人工神经网络将氧化微晶蜡的酸值、酯值及微晶蜡的氧化条件（反应温度、空气流量和反应时间）进行关联，建立了微晶蜡非催化氧化的酸值、酯值的神经网络模型，并用该模型预测了反应条件对微晶蜡氧化反应过程的影响。结果表明，该模型不但具有较高的计算精度，而且具有满意的预测能力。     

Neural networks in the study of the brain

Neural networks are models of the brain and have been used within Artificial Intelligence to provide alternative explanations to the symbolic explanations of cognition in which one assumes that an intelligent system has certain explicit representations of some aspect of the world and uses these in intelligent behavior. Obviously, if neural networks are indeed good models of the brain, and give a satisfactory account of cognition, then they could be a valuable tool to neuroscientists. This article gives a brief overview of the various neural network models, and critically reviews their status as models of the brain and of cognition.     

Connectionist networks learn to transmit chaos

Evidence presented in the preceding paper indicates that the activity of some neurons during the generation of coordinated motor patterns may be attributable to chaos. Because even "simple" biological systems are difficult to control, we have used connectionist networks in order to inquire into the question of whether a chaotic signal originating in one part of the nervous system can be learned and transmitted by another. We have examined a number of different architectures, and report here the findings for a simple network consisting of one input unit, four hidden units, and one output unit. During training sessions, the input of the circuit was given analog values of either the 3.60 or 3.95 logistic equation, or of one variable of the three-variable R?ssler attractor. The backpropagated error in the learning algorithm was a function of the difference between the input value and the output at each iteration. Iterations involving small changes in analog value resulted in good similarity between the input and output signals, but little learning occurred because of the small error propagated back to the synapses. With larger differences in the analog values (and larger feedback error) at each iteration, we found that networks learned to transmit different chaotic attractors. Once the network learned one input, it could transmit another without changing the synapses. Increasing the number of hidden units increased the rate of learning.(ABSTRACT TRUNCATED AT 250 WORDS)     

Gas-to-liquid permeation in silicon-containing, crosslinked, glassy copotymers of methyl methacrylate

Copolymers of methyl methacrylate with disiloxane derivatives have been proposed as biomaterials for contact lens applications. Although glassy, these copolymers exhibit high oxygen permeability and adequate wettability so that they can be used for manufacture of hard, extended wear lenses. CrossHnked copolymers of poly(methyl methacrylate-co-1,3-bis(methacryloxymethyl)-1,1,3,3-tetramethyl-disiloxane), P(MMA-co-BMTDS), containing from 0.085 to 0.53 mole fraction of BMTDS were prepared and tested for oxygen permeation using a novel apparatus which simulates the atmosphere/lens/cornea conditions. The gas-to-liquid dissolved oxygen permeability, P_gd, was determined and it was found to increase with BMTDS content Permeability values for P(MMA-co-BMTDS) at 34°C were significantly higher than for pure homopolymer PMMA, although these copolymers were glassy at this temperature. The increased oxygen permeation was attributed to increased oxygen solubility in the copolymers due to the presence of the -Si-0-bonds.     

Vasculogenic mimicry has no prognostic significance in pT3 and pT4 cutaneous melanoma

The concept of vasculogenic mimicry has been introduced to define periodic acid-Schiff (PAS)-positive channels and loops lined by tumor cells, instead of endothelium, able to contribute to microcirculation in uveal melanomas. Previous studies have shown that the PAS-positive patterns are associated with a poor prognosis in uveal melanoma. The aim of the current study was to investigate whether vasculogenic mimicry has a prognostic impact in pT3 and pT4 cutaneous melanoma. Fifteen patients with pT3 and pT4 cutaneous melanoma who did not experience progression after 10 years of follow-up and 30 matched controls who underwent progression were selected. Tumor sections were stained with PAS reaction, omitting the nuclear counterstaining. For immunohistochemistry, sections were stained with CD31, CD105 (endoglin), and laminin. Differences in the distribution of the PAS-positive patterns and a series of clinicopathological variables were evaluated by the Pearson chi(2) and Mann-Whitney U tests. We observed PAS-positive linear sheets, arcs, elliptical loops, and networks encircling roundish to oval aggregates of melanoma cells. The overall distribution of the PAS-positive patterns did not match with the blood microvessels' architecture as detected by immunohistochemical analysis. No statistically significant differences in the distribution of PAS-positive patterns were found between cases and controls. The presence of a parallel pattern correlated significantly with thickness (P = 0.04), whereas an inverse correlation was found with vessel area (P = 0.05). In conclusion, our results suggest that there is a mismatch between vasculogenic mimicry and tumor angiogenesis and do not support any prognostic role of vasculogenic mimicry in thick cutaneous melanoma.     

Automated Recognition of Lateral from PA Chest Radiographs: Saving Seconds in a PACS Environment

Images acquired in a two-view digital chest examination are frequently not electronically distinguishable. As a result the lateral and posterioanterio (PA) images are often improperly positioned on a PACS work station. A series of 1998 chest radiographs (999 lateral, 999 PA or AP) were used to develop a neural network classifier. The images were down-sampled to 16 × 16 matrices, and a feed-forward neural network was trained and tested using the leave-one-out method. Using five nodes in the hidden layer, the neural network correctly identified 987 of the 999 test cases (98.8%) (average of six runs). The simple architecture and speed of this technique suggests that it would be a useful addition to PACS work station software. The accumulated time saved by correctly positioning the lateral and PA chest images on the work station monitors in accordance with each radiologists hanging protocols was estimated to be about 1 week of radiologist time per year.     

An intelligent system for diagnosis of the heart valve diseases with wavelet packet neural networks

In this paper, an intelligent system is presented for interpretation of the Doppler signals of the heart valve diseases based on the pattern recognition. This paper especially deals with combination of the feature extraction and classification from measured Doppler signal waveforms at the heart valve using the Doppler Ultrasound. Because of this, a wavelet packet neural network model developed by us is used. The model consists of two layers: wavelet and multi-layer perceptron. The wavelet layer is used for adaptive feature extraction in the time-frequency domain and is composed of wavelet packet decomposition and wavelet packet entropy. The multi-layer perceptron used for classification is a feed-forward neural network. The performance of the developed system has been evaluated in 215 samples. The test results showed that this system was effective in detecting Doppler heart sounds. The correct classification rate was about 94% for abnormal and normal subjects.     

An integrative approach to CTL epitope prediction: a combined algorithm integrating MHC class I binding, TAP transport efficiency, and proteasomal cleavage predictions

Reverse immunogenetic approaches attempt to optimize the selection of candidate epitopes, and thus minimize the experimental effort needed to identify new epitopes. When predicting cytotoxic T cell epitopes, the main focus has been on the highly specific MHC class I binding event. Methods have also been developed for predicting the antigen-processing steps preceding MHC class I binding, including proteasomal cleavage and transporter associated with antigen processing (TAP) transport efficiency. Here, we use a dataset obtained from the SYFPEITHI database to show that a method integrating predictions of MHC class I binding affinity, TAP transport efficiency, and C-terminal proteasomal cleavage outperforms any of the individual methods. Using an independent evaluation dataset of HIV epitopes from the Los Alamos database, the validity of the integrated method is confirmed. The performance of the integrated method is found to be significantly higher than that of the two publicly available prediction methods BIMAS and SYFPEITHI. To identify 85% of the epitopes in the HIV dataset, 9% and 10% of all possible nonamers in the HIV proteins must be tested when using the BIMAS and SYFPEITHI methods, respectively, for the selection of candidate epitopes. This number is reduced to 7% when using the integrated method. In practical terms, this means that the experimental effort needed to identify an epitope in a hypothetical protein with 85% probability is reduced by 20-30% when using the integrated method.The method is available at http://www.cbs.dtu.dk/services/NetCTL. Supplementary material is available at http://www.cbs.dtu.dk/suppl/immunology/CTL.php.     

Dialysate-side urea kinetics. Neural network predicts dialysis dose during dialysis

Determination of the adequacy of dialysis is a routine but crucial procedure in patient evaluation. The total dialysis dose, expressed as Kt/V, has been widely recognised to be a major determinant of morbidity and mortality in haemodialysed patients. Many different factors influence the correct determination of Kt/V, such as urea sequestration in different body compartments, access and cardiopulmonary recirculation. These factors are responsible for urea rebound after the end of the haemodialysis session, causing poor Kt/V estimation. There are many techniques that try to overcome this problem. Some of them use analysis of blood-side urea samples, and in recent years, on-line urea monitors have become available to calculate haemodialysis dose from dialysate-side urea kinetics. All these methods require waiting until the end of the session to calculate the Kt/V dose. In this work, a neural network (NN) method is presented for early prediction of the Kt/V dose. Two different portions of the dialysate urea concentration-time profile (provided by an on-line urea minitor) were analysed: the entire curve A and the first half B, using an NN to predict the Kt/V and compare this with that provided by the monitor. The NN was able to predict Kt/V is the middle of the 4h session (B data) without a significant increase in the percentage error (B data: 6.69%±2.46%; A data: 5.58%±8.77%, mean±SD) compared with the monitor Kt/V.